Engineering AI and Security Tabs — Documenting What Agents and CI Actually Run
The engineering hub already had an infrastructure map, a delivery workflow page, and a portfolio case study with a live CI snapshot. Visitors evaluating platform work could see topology and Kanban — but two stories lived elsewhere: how coding agents orient in the repo without paid LLM APIs, and how every push hits Gitleaks and Trivy before npm ci runs.
The low-RAM Woodpecker article covers why the security lane is serialized. The AI context doc in docs/AI_CONTEXT.md covers Graphify and doctor/repair for developers. This article covers why those became first-class engineering tabs at /engineering/ai and /engineering/security — and how the pages stay thin by pulling from data modules instead of hardcoding copy in JSX.
The portfolio case study is meta — it documents the pipeline that ships this site. Infrastructure is spatial: clusters, Traefik, Forgejo, Netdata. Delivery is temporal: Plane tickets, PORT-N branches, Kanban states. AI context and security gates are cross-cutting concerns that apply to every repo on the platform, not just the portfolio.